Comparison of nanoindentation results obtained with
Berkovich and Cube Corner indenters
Thomas Chudoba, Patrick Schwaller, R. Rabe, J.-M. Breguet, Johann Michler
To cite this version:
Thomas Chudoba, Patrick Schwaller, R. Rabe, J.-M. Breguet, Johann Michler. Comparison
of nanoindentation results obtained with Berkovich and Cube Corner indenters. Philosophical
Magazine, Taylor & Francis, 2006, 86 (33-35), pp.5265-5283. .
HAL Id: hal-00513706
https://hal.archives-ouvertes.fr/hal-00513706
Submitted on 1 Sep 2010
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Philosophical Magazine & Philosophical Magazine Letters
rP
Fo
Journal:
rR
ee
Comparison of nanoindentation results obtained with
Berkovich and Cube Corner indenters
Philosophical Magazine & Philosophical Magazine Letters
Manuscript ID:
TPHM-05-Nov-0528.R1
Journal Selection:
Philosophical Magazine
Chudoba, Thomas; ASMEC GmbH
Schwaller, Patrick; Empa, Materials Technology
Rabe, R.; EPF Lausanne, Institut de Production et Robotique
Breguet, J.-M.; EPF Lausanne, Institut de Production et Robotique
Michler, Johann; EMPA
w
ie
Complete List of Authors:
06-Feb-2006
ev
Date Submitted by the
Author:
Keywords:
Cube corner indenter
ly
On
Keywords (user supplied):
nanoindentation, hardness, indentation hardness, indentation
testing, mechanical behaviour, mechanical testing
http://mc.manuscriptcentral.com/pm-pml
Page 1 of 34
Comparison of nanoindentation results obtained with
Berkovich and Cube Corner indenters
T. Chudoba*1, P. Schwaller2, R. Rabe2, 3, J.-M. Breguet3, J. Michler2
1
ASMEC GmbH, D-01454 Radeberg, Germany
E-Mail: [email protected]
2
Fo
Materials Technology Laboratory, Empa, Materials Science and Technology, CH- 3602
Thun, Switzerland
3
rP
Institut de Production et Robotique, IPR-LSRO, EPF Lausanne, CH-1015 Lausanne-EPFL,
Switzerland
Abstract
rR
ee
There is increasing interest in using sharp cube corner indenters in nanoindentation
experiments to study plastic properties. In combination with finite element methods it is for
ev
instance possible to extract stress-strain curves from load-displacement curves measured with
differently shaped pyramidal indenters. Another example is the fracture toughness of
iew
coatings, which can be studied using cracks produced during indentation with cube corner
tips. We have carried out indentation experiments with Berkovich and cube corner indenters
on eight different materials with different mechanical properties. To gain information about
On
the formation of pile-up and cracks, indentation experiments with cube corner indenter were
performed inside a Scanning Electron Microscope (SEM) using a custom-built SEMMicroindenter. The results show that reliable hardness and modulus values can be measured
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
using cube corner indenters. However, the fit range of the unloading curve has a much bigger
influence on the results for the cube corner than for the Berkovich tip. The unloading curves
of a cube corner measurement should therefore be carefully inspected to determine the region
of smooth curvature and the unloading fit range chosen warily. Comparison of the modulus
results shows that there is no significant difference between cube corner and Berkovich
measurements. Also for hardness, no fundamental difference is observed for most of the
investigated materials. Exceptions are materials like silicon nitride, cemented carbide or
glassy carbon, where a clear difference to the hardness reference value has been observed
although the modulus difference is not pronounced.
Chudoba et al., page 1 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
* Corresponding author: [email protected]
Keywords:
elastic properties, hardness, nanoindentation
iew
ev
rR
ee
rP
Fo
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chudoba et al., page 2 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 2 of 34
Page 3 of 34
1. Introduction
Nanoindentation is nowadays an established tool for the measurement of hardness and
Young’s modulus of surfaces and thin films. In the thin film community the use of three-sided
Berkovich pyramids as indenters became a standard in contrast to four-sided Vickers
pyramids that are mostly used in conventional hardness tests. The reason for this preference is
the possibility to prepare sharper tips with smaller tip rounding. Vickers indenters may show
the so called “chisel edge” where the four faces don’t meet each other in one single point.
Fo
This results in asymmetrical impressions, especially for small indents, and complicates the
determination of an accurate area function. Another indenter with the shape of a three-sided
rP
pyramid is the cube corner. A diamond cube is simply used in such a way that the indentation
direction is along the space diagonal which connects opposite corners. This indenter with a
face angle (angle between surface normal of sample and indenter face) of 35.3° is much
ee
sharper than a Berkovich pyramid with a face angle of 65.3°. Radial cracks, which normally
start at the corners of the pyramid, are therefore easier to induce. These cracks can be used for
rR
the investigation of the fracture toughness of coatings (see [1, 2]). However, for the
measurement of hardness and Young’s modulus cracks are undesirable since they may cause
measurement errors.
ev
In recent years there has been, however, a growing interest to use cube corner indenters as an
alternative indenter type to study plastic properties. Due to the self-similarity of pyramidal
iew
indenters it is not possible to realize different strain states with one and the same tip. Beside
the very first depth range where the tip rounding plays a role, the stress field is growing but
keeping the same stress ratio with the increasing contact area. This disadvantage can be
On
overcome by using tips with different face angles. Stauss et al. and Bucaille et al. [3,4,5] have
shown, by using Berkovich and cube corner indenters combined with finite element
calculations the same stress-strain values as in tensile tests can be obtained with this method.
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
Chollacoop et al. [6] have applied a similar method with Berkovich and 60° cone tips. Field et
al. performed indentations on fused silica and glassy carbon with different tip geometries
including Berkovich and cube corner for the determination of the fracture toughness [7]. In
this study the hardness and modulus values were however not determined from the
indentation data but taken from the literature.
Little work has been done, however, to use cube corner indenters to measure Young’s
modulus and hardness in the same way as Berkovich indenters and to check the applicability
of the Oliver and Pharr analysis method [8]. Generally it is expected that the much sharper
Chudoba et al., page 3 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
cube corner tip deforms the material more by a cutting mechanism and results in a stronger
pile-up. Such an upward flow of material can be described the slip-line theory, developed
originally in two dimensions by Hill [9]. By comparison, the plastic zone of a blunt Berkovich
tip is constrained by the surrounding material and upwards flow of material is only possible if
the radial expansion of the plastic zone is big enough. It is assumed that the displaced volume
is taken up entirely by elastic strains within the specimen material. This deformation
mechanism can be described by the expanding cavity models of Marsh [10] and Johnson [11].
The transition range between the two models is not well defined and depends on the material
Fo
and the friction coefficient between indenter and sample. Therefore it is possible that
Berkovich as well as cube corner indents can be described with the same model. Zhang and
rP
Sakai [12] investigated geometrical effects of pyramidal indenters with different face angles
on the elastoplastic contact behaviors of ceramics and metals. They found different hardness
results depending on the face angle; however their results cannot be compared to ours because
ee
they calculated the hardness values with a model which assumes a quadratic behavior of the
unloading curves which we did not observe for any of our investigated materials. Further they
rR
tried to derive a face angle independent “true” hardness value that was related to the flow
stress by a constraint factor C. Indeed, H=C*Y is the well known Tabor relation for metals
ev
[13] with a constraint factor of about 3 for a representative strain of 8%. According to this
model hardness should only weakly depend on the face angle.
iew
We have carried out indentation experiments with Berkovich and cube corner indenters on
eight different homogeneous materials with two different commercial nanoindentation
instruments (UMIS-2000, CSIRO, Australia and Nanoindenter XP, MTS, USA). A large load
range between 0.3 mN and 500 mN was used to obtain depth dependent results. Fused silica
On
and sapphire were used as reference materials for a careful determination of the compliance of
the two instruments and the indenter area functions. Finally hardness and modulus results
were calculated according to ISO 14577 part 1 [14] and compared. The results are discussed
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
in relation to reference values obtained with other measurement methods. To gain information
about the formation of pile-up and/or crack formation with cube corner indenters, additional
indentation experiments were performed inside a Scanning Electron Microscope (SEM) using
a custom-built SEM-Microindenter (for details see [15]) using the same tip as for the MTS
Nanoindenter XP measurements.
Chudoba et al., page 4 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 4 of 34
Page 5 of 34
2. Experimental
2.1. Material selection
Eight materials covering a wide range of mechanical properties were investigated in this
study. The reference values for the hardness and the Young’s modulus of these materials are
listed in table 1. Most of the samples were also used in the European project “Certified
Reference Materials for Depth Sensing Indentation Instrumentation” (DESIRED) [16] which
was finished in February 2004. The Ni sample was already investigated and used as substrate
in the European project INDICOAT [17]. A special ultra-fine grain cemented carbide
Fo
(hardmetal) sample was delivered by the company Widia, Germany. In contrast to
conventional cemented carbide the grain size of the WC particle was in the order of 100 nm
rP
and the cobalt content was very low. The distribution of the WC particles in this sample was
highly homogeneous and its Young’s modulus close to the value of pure WC.
ee
The conventional Vickers hardness and the Young’s modulus of the materials were
determined independent of the nanoindentation experiments as reference values. The results
rR
are given in table 1. Vickers hardness tests have been carried out by trained operators at the
Federal Institute of Materials Research and Testing (BAM) in Berlin with a Reichert-Jung
Micro-Duromat 4000. The magnification of the optical system was 1000 x. Both diagonals of
ev
the indents were measured and average length was determined. A load of 500 mN was applied
in 30 s and the hold period at maximum load was 12 s. The instrument is part of a quality
iew
control system according to ISO 17025. The calibration is checked every half-year according
to DIN 51220 with the help of a hardness reference block with 315 ± 10 HV 0.1 and 333 ± 24
HV 0.015.
Ten measurements were done on every sample and the Vickers hardness results were
On
averaged. The equivalent HIT value was estimated using the informative relationship given in
ISO 14577: HIT = HV/0.0945. Table 1 gives the hardness HIT and additionally the standard
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
deviation of the hardness results σH.
The Young’s modulus of fused silica was measured within the INDICOAT project [17] with
different methods (Laser Surface Acoustic Wave, Acoustic Microscopy, Brillouin scattering).
The moduli of the Au-Cu alloy (containing 18% Cu), Si3N4 and cemented carbide (hardmetal)
were measured with surface acoustic waves at the BAM and the Fraunhofer Institute IWS in
Dresden, Germany. The moduli of W and Ni are literature data [18]. The modulus of Sigradur
glassy carbon is an average value from indentation measurements done with different
instruments and loads. The modulus of sapphire was calculated as a weighted elastic average
on basis of the single crystal elastic constants for hexagonal symmetry. The sapphire surface
Chudoba et al., page 5 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
had a (0001) orientation. Perpendicular to this basal plain in the direction of indentation is the
absolute modulus maximum with E = 461 GPa. Directions within the (0001) plain have a
modulus of about 380 GPa which was confirmed by surface acoustic wave measurements.
The modulus average over the whole space is 404.7 GPa. The indentation direction with the
highest modulus is higher weighted than the other directions and the effective modulus is
therefore slightly higher than the average with about 420 GPa.
Fo
2.2. Nanoindentation measurements
The experiments were carried out with an UMIS-2000 Nanoindenter (CSIRO, Australia) at
rP
the Technical University of Chemnitz, Germany and a Nanoindenter XP (MTS, Knoxville,
USA) at EMPA in Thun, Switzerland. A measurement cycle consisted of a loading segment,
ee
followed by a first hold period at constant maximum load to reduce creep effects which could
influence the shape of the unloading curve. The unloading segment was followed by a second
hold period at about 10% of the maximum load for the determination of the thermal drift of
rR
the instrument. Different loading and unloading times were used depending on the maximum
load and the instrument. A typical total measurement time for the UMIS was 400 s and for the
ev
Nanoindenter XP 140 s. The loading time of the UMIS depends on the number of data points
and cannot be varied markedly. In contrast to the Nanoindenter XP, which used a linear force
iew
increase and decrease over time, the force steps followed a quadratic rule. The hold period at
maximum load was set to 30 s for the XP and between 30 s and 60 s for the UMIS. The
second hold period at the end of the measurement cycle was 60 s for the XP and 45 s for the
UMIS. Only the last 30 s were used for both instruments to apply a linear fit and to determine
On
the thermal drift rate. This was possible with an accuracy of 0.003 nm/s for the UMIS, and
0.005 nm/s for the XP, respectively.
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The true surface position was determined by a correction procedure that uses a fit of the first
displacement data points after contact for a back-extrapolation to zero depth. The fit was done
assuming a Hertzian contact for the first part of the loading segment (typically below 30 nm
depth). The zero point correction is necessary for the UMIS because the instrument software
only recognizes the surface if the contact force exceeds about 5-10 µN. The deformation at
this force and the influence of surface roughness can result in a zero shift of more than 3 nm.
This would influence the results markedly when the absolute deformation was small. The
determination of the zero position of the displacement axis could also be improved by this
Chudoba et al., page 6 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 6 of 34
Page 7 of 34
method for the XP although it uses another principle for the surface detection. More details
can be found in [19].
Between 8 and 10 different loads within a range between 300 µN and 500 mN were used to
measure the reference materials fused silica and sapphire. For the other materials at least 5
different loads (5 mN, 30 mN, 100 mN, 300 mN and 500 mN) were applied with the XP.
UMIS measurements were carried out with some additional loads of 300 µN, 1 mN, 3 mN and
50 mN. 10 indentations separated by a distance of 50 µm were made for each load on every
sample and average load-displacement curves were calculated. Finally, the averaged depth
Fo
values were corrected for the instrument compliance at the corresponding load and analyzed.
This is in contrast to the usual analysis of every single curve and reduces the uncertainty of
rP
the data due to the averaging. Additionally, it reduces the amount of data to be handled. A
pop-in event would be smeared out in such an average curve but this is not a problem here
because such effects were not included in this work. Chudoba has shown in previous work
ee
[17] that the results of this procedure agree well with the hardness or modulus average of the
results from every single curve if the curves are smooth.
rR
The data analysis for both instruments, including all corrections, was done with the
instrument independent software IndentAnalyser® (ASMEC GmbH, Radeberg, Germany).
iew
ev
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
Chudoba et al., page 7 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
2.3. Determination of the instrument compliance
An accurate knowledge of the instrument compliance is indispensable for calculating reliable
hardness and Young’s modulus values from load-displacement curves, especially at higher
loads. We determined the instrument stiffness by a comparison of the area functions obtained
using data of two different materials (fused silica and sapphire). The measurements were done
using a Berkovich tip and the same stiffness function was used for the evaluation of cube
corner data. Because the influence of the instrument stiffness becomes more important for
samples with a high modulus, it is recommended to include a material with a high modulus
Fo
(sapphire) for the instrument stiffness determination. In the following, the procedure for the
determination of the instrument stiffness is described in detail:
rP
First 10 load-displacement curves for fused silica and sapphire with different nominal
maximum loads (500, 300, 100, 50, 30, 10, 5, 3, 1, 0.5, and 0.3 mN, respectively) were
measured. From the average curves the effective maximum load F, the unloading stiffness S,
ee
the maximum depth hmax, the depth after unloading h0, and the unloading exponent m were
determined. It is important to note that these quantities (with the exception of m) were
rR
obtained from the load-displacement curve without a correction of instrument compliance.
The instrument compliance Cf was determined such that the area functions obtained from both
ev
reference materials agree. This was done under the assumption that the compliance Cf is the
only parameter responsible for the difference in the area function. The contact depth hc is
related to Cf according to
h c = hmax −
iew
F
− ε(m) ⋅ F ⋅ (C − C f )
Sf
(1)
On
ε describes the ratio between the elastic deformation above and below the contact area. There
is an unequivocal relation between ε and the unloading exponent m (see for instance [20]).
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 8 of 34
ε=1 belongs to m=1 (flat punch indenter geometry); ε=0.75 belongs to m=1.5 (sphere) and
ε=0.727 to m=2 (cone). For real materials after plastic deformation ε may vary according to a
complicated function between 0.7 and 0.8.
The square root of the contact area AC is given by
AC =
π 1
1
(1 + ur )
2 Er C − C f
(2)
where Er is the reduced modulus
Chudoba et al., page 8 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 9 of 34
2
1− ν s
1− νi
1
=
+
Er
Ei
Es
2
(3)
ur is the radial displacement correction;
ur =
h
(1 − 2ν )(1 + ν ) F
cos(arctan( 0 ))
2
2
a
Er ⋅ a
(4)
h0 is the residual depth, and a is the equivalent contact radius
Fo
a=
Ac
π
(5)
The additional radial displacement correction which is not included in ISO standard 14577
rP
[14], and allows for elastic displacement within the surface plane. Although the lateral elastic
displacement in indentation experiments has already been estimated by Johnson [11] and
others it is normally not considered on the assumption that it can be neglected. In contrast to
ee
Johnson’s result for fully elastic deformation with a spherical indenter, equation (4) is also
valid for additional plastic deformation.
rR
Sometimes an additional geometrical factor 1/β is incorporated in eq. (2) in order to account
for the non-rotational symmetric shape of the pyramids. However all calculations in the
ev
literature (for instance [21] or [22]) were done for the fully elastic case and gave values
between 1.02 and 1.14 for the Berkovich indenter. For elastoplastic deformations the infinite
iew
elastic stresses at edges and tip will be reduced drastically by the plastic displacement of
material. It is well known that the plastic zone of pyramidal indentations is approximately
rotationally symmetric (see for instance [11]) and that the elastic field after plastic
deformation has a similar shape. For this reasons, no value is given or recommended for beta
On
in the standard ISO 14577 and it is indeed better to use beta = 1 for pyramidal tips.
The Young’s modulus E and Poisson’s ratio values ν of both indenter (index i) and sample
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
(index s) were assumed to be known. We took the following values for the calculations: For
diamond (E = 1140 GPa, ν = 0.07), for fused silica (E = 72 GPa, ν = 0.17), and for sapphire
(E = 420 GPa, ν = 0.24), respectively.
To begin with, the instrument stiffness was set to infinite. As already mentioned, the
instrument stiffness has a larger influence on the measurement of a material with a higher
modulus. Therefore the √A(h)-curve for the stiffer material lies below the curve of the other
and the resulting square root of the contact area, according to equation (2), changes more with
varying instrument stiffness than for the less stiff material. At a certain instrument stiffness
value both curves meet each other. The procedure was therefore to decrease the instrument
Chudoba et al., page 9 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
stiffness stepwise decreased from 106 mN/µm until both area functions agreed as closely as
possible. This was done for both area functions with the formulas given above. A relatively
complicated iterative procedure is repeated six times to get, finally, a force dependent
stiffness function. Stiffness results for forces below 30 mN were ignored because the accuracy
of the stiffness in this range is normally not good enough. Below 30 mN constant instrument
stiffness was assumed. The mean instrument compliance (inverse stiffness) of the UMIS-2000
was 0.21 nm/mN. It was slightly increased towards smaller forces up to 0.286 nm/mN. The
mean compliance of the Nanoindenter XP was 0.15 nm/mN. It increased to 0.16 nm/mN at
30mN.
iew
ev
rR
ee
rP
Fo
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 10 of 34
Chudoba et al., page 10 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 11 of 34
2.4. Determination of indenter area function
The area function A(hc) for the Berkovich indentation tip was calculated with the same
measurement data of fused silica and sapphire, and with the same formulas (see above) as was
used for the stiffness calibration. The compliance function as determined before, was used for
all calculations. The calculations were done according to the ISO 14577 standard but allowed,
in addition, a variable ε factor and a radial displacement correction ur as mentioned above.
This is a refinement of the ISO standard method and it improves the accuracy of the analysis
by some percent. The influence of these corrections depends on the material. They have the
Fo
largest influence on fused silica. The increase in the value of the square root of AC calculated
by equation (2) is about 5% for the Berkovich indenter. This results in a 5% reduction in the
rP
modulus and a 10% reduction in the hardness, because hardness scales with 1/AC. The
increase of √AC is the smallest for Ni at only 0.25%. Fused silica has (with m = 1.2 for
Berkovich) the smallest unloading exponent of all investigated materials. The standard
ee
epsilon value of 0.75 is only absolutely correct for an unloading exponent of 1.5.
An
unloading exponent of 1.2 corresponds to an epsilon value of 0.8 and increases the square root
rR
of AC for fused silica by 2.3%. In contrast, sapphire has an unloading exponent of about 1.5
and the correction has no influence for this material.
ev
For the Berkovich measurements, a fit range of the unloading curve between 98% and 40% of
the maximum load Fmax was used. The same values for E and ν as indicated in the previous
iew
section were used. It turned out that these parameters did not yield good results for the
calibration of the cube corner tip with fused silica. In contrast to the Berkovich tip, for the
cube corner tip, it could be seen that a kink occurs in the unloading curves at about 50%
unloading. A fit with a power function gave an unloading exponent of about 1.4 if the fit
On
range was restricted to the data above the kink. Otherwise the unloading exponent was two or
higher. To avoid this kink influencing the accuracy of the fit of the unloading curve in the
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
evaluation of cube corner data, a reduced range, between 98% and 50% of the unloading
curve, was used for the curve fit.
A second aspect to consider was the reference value of the Young’s modulus for the sapphire
sample. In contrast to the other materials investigated in this work, the single-crystalline
sapphire sample is elastically anisotropic and has six independent elastic constants (see for
instance Wachtmann [23] and Vlassak [24]). Due to the different indentation shapes induced
by plastic deformation with Berkovich and cube corner indenters, the contributions of the
various crystallographic directions around the indent to the absolute elastic response may be
different. We mentioned that the influence of the stiffest direction perpendicular to the (0001)
Chudoba et al., page 11 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
surface plane is less pronounced for the sharper cube corner tip. For cube corner data the best
calibration was obtained by using a modulus of 408 GPa instead of the 420 GPa value found
for the Berkovich indenter. This is closer to the isotropic average of 405 GPa.
The tip rounding of the UMIS Berkovich, a much used tip, was much bigger than that of the
Nanoindenter XP. The tip radius of the UMIS Berkovich was determined to be about 0.84 µm
while the radius of the XP Berkovich and the cube corner indenter was about 0.25 µm and
less than 0.15 µm, respectively. The radius was obtained from an elastic fit of the loaddisplacement curve in the very low load range using the Hertzian model. The large tip radius
Fo
of the UMIS tip has an effect on the hardness results at small loads, as is shown below. The
different tip radius also influences the maximum indentation depth. For example sapphire
rP
measurements at 10 mN with the UMIS Berkovich were purely elastic with a maximum depth
of 86 nm. The same measurements with the XP Berkovich and cube corner indenters were
plastic and reached indentation depths of 141 nm and 336 nm, respectively. In this example
ee
the sharper cube corner tip gives rise to a 2.4 times larger indentation depth. This illustrates
the much higher strain induced with the cube corner indenter.
iew
ev
rR
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 12 of 34
Chudoba et al., page 12 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 13 of 34
3. Results and Discussion
3.1. Influence of the unloading fit range
Figure 1 shows the Young’s modulus values obtained from Berkovich (squares) and cube
corner (triangles) data measured for fused silica (Fig. 1a), sapphire (Fig. 1b) and Sigradur
glassy carbon (Fig. 1c). The maximum load for fused silica and glassy carbon was 500 mN,
for sapphire 100 mN. Higher applied loads for sapphire led to crack formation with large popin events in the load displacement curves. Different ranges of the unloading curve have been
Fo
fitted. The upper fit limit was set to 100% of the maximum load Fmax, the lower limit varied
between 80% and 10% of Fmax. A fit range of 90% from Fmax represents a fit from 100% down
rP
to 10% of Fmax. Therefore the fit range increases from left to right in Fig. 1. The horizontal
solid lines indicate the reference Young’s modulus values (see Table 1) obtained with other
methods. As can be deduced from the graphs of Fig. 1 the modulus values calculated from
ee
Berkovich measurements are almost unaffected by a variation of the unloading fit range and
the experimental values are in fair agreement with the reference values. For cube corner data
rR
however, a strong influence of the fit range is evident. For fused silica and sapphire an
increase in the modulus result is observed with increasing fit range. Note that for both
ev
materials the values closest to the reference are obtained for a lower fit limit between 50 and
60% of Fmax, a range which was used in the following for the data evaluation of all loads. For
glassy carbon the modulus decreases with increasing fit range. Only for a fit between 100%
iew
and 80% Fmax can the reference value be obtained. The reason for this can be found by a
closer look at the unloading curve (not shown here). There is a distinct kink in the cube corner
curve after unloading to 80% Fmax (F = 400 mN). Therefore only the upper 20% data were
On
used for the fit of the unloading curve of glassy carbon in the following data analysis. The
reason for the kink is not clear. Possibly it is connected with first cracking. Cracks parallel to
the surface can be seen in the SEM images (see Fig. 3).
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
A similar but less pronounced kink can be seen in the unloading curves of fused silica at 50%
Fmax as mentioned previously. These kinks were nearly independent of the maximum load
down to about 10mN. The results show that the unloading fit range is a critical issue for cube
corner data analysis. The unloading curves should be carefully inspected for a smooth
curvature. Instead of the recommended 80% fit range for Berkovich or Vickers indenters in
[14], a fit range of less than 50% should be preferentially used to get comparable results.
Chudoba et al., page 13 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
3.2. Modulus results
The results of all measurements are summarized in Table 1. All values for one material
obtained at different loads were averaged. Hardness results below 40 nm contact depth,
however, were not considered in the average to exclude the influence of tip rounding. The
indicated errors are the standard deviations obtained from the different loads without
considering the scatter of ten measurements performed at one load. This standard deviation
represents therefore the homogeneity of the materials over the depth.
Fig. 2 shows the depth dependent modulus results of all investigated materials and all three
Fo
indenters in relation to the reference values. Cube corner results are depicted by triangles,
Berkovich results from Nanoindenter XP are depicted by squares and Berkovich results from
rP
UMIS-2000 are depicted by spheres. As expected, the results of the reference materials fused
silica and sapphire agree well with the reference data because they were used for the
calibration of instrument compliance and area function. The scatter of the sapphire data is
ee
slightly higher and some of the cube corner results for sapphire are too low. However, a trend
in the deviations cannot be seen. This shows that the two materials are suitable for the
rR
calibration of the area function of a cube corner tip in the used load range.
The same XP cube corner and Berkovich tips were used for indentation experiments inside a
ev
Scanning Electron Microscope (SEM) using a custom-built SEM-Microindenter. More details
about this instrument can be found in [15]. Fig. 3 shows indentations with cube corner
iew
indenter into sapphire, Ni, W and glassy carbon. Details of these images will be discussed
below. Berkovich indentations inside the SEM were done for fused silica and sapphire loads
above 700mN (not shown here). No cracks or chipping could be observed in fused silica at
736 mN while cracks along the edges could be seen for a higher force of 1226 mN. In
On
sapphire, the surface around the Berkovich indents was surprisingly smooth. Only cracks
starting at the corners could be seen. A cube corner indent into sapphire is shown in Fig. 3a.
As expected, the fracture effects are much more pronounced for the cube corner indenter than
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 14 of 34
for the Berkovich tip.
It can be seen in Fig. 2 that there is a good agreement of the modulus results with the
reference values for all three indenters for W, Au-Cu alloy and cemented carbide. A clear
depth dependence of the result cannot be observed. Pile-up effects do not seem to influence
the modulus results markedly. The scatter of the cube corner results is higher than for the
Berkovich indenter; however there is no indication that the analysis method cannot be applied
for cube corner indentations. The good agreement of the cemented carbide results with the
reference value, being by far the material with the highest modulus, indicates that the
Chudoba et al., page 14 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 15 of 34
instrument stiffness was correctly determined. The Si3N4 modulus results of the Berkovich
indenter agree well with the reference value of 304 GPa, but the cube corner indenter yields
incorrect results. The obtained modulus is significantly lower than the reference value and the
discrepancy increases with decreasing load. We attribute this finding to the higher sensitivity
of the sharper cube corner tip to the grain size and the brittleness of the ceramic material.
Probably the results are affected by cracking effects.
The modulus results for glassy carbon with its high elasticity (the energy release is about
90%) are close to the reference value. The difference for the cube corner indenter is slightly
Fo
higher but it is not significant. The Ni results for the three indenters agree well, however they
are all higher than the reference value. This can be attributed to the marked pile-up effect for
rP
this material (see fig. 3d). The contact area derived from the indentation data is therefore too
small. W also shows a pile-up effect, however it is less pronounced and seems not to
influence the modulus results. Fig. 3g and h shows the crack development in glassy carbon.
ee
Radial cracks can be seen at the corners of the indenter and lateral cracks have developed
inside the indent, parallel to the surface. The inwards bowing of the edges after unloading is
rR
due to the high elasticity of the material and indicates a sink-in effect. Usually a sink-in effect
is connected with an overestimation of the contact area and therefore an underestimation of
ev
the hardness. However, the hardness results for this material (see below) are rather too high
which contradicts this interpretation. The contact area cannot be calculated accurately enough
iew
from the load-displacement curve alone as this material does not appear to behave in a manner
consistent with conventional contact mechanics.
3.3. Hardness results
On
The depth dependent hardness results for all investigated materials are shown in Fig. 4.
Several effects can be observed. The hardness results are generally not as constant as the
modulus results for all tips. The fused silica results obtained with the UMIS Berkovich
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
decreases markedly towards smaller contact depths. This is caused by the large tip radius of
0.84 µm. Measurements at 0.3 mN (first data point) are fully elastic for this tip and the
hardness value is only the contact pressure for this load. At higher loads the elastic-plastic
transition takes place and only above about 2 mN (third data point) is a fully developed plastic
zone established. The cube corner results for fused silica show a slight minimum at medium
loads between 5 mN and 10 mN (200 nm – 800 nm depth). The reason for this effect is
unknown. Generally the indentation hardness results for fused silica are about 1 GPa higher
than the reference values. The difference is less pronounced for smaller loads and sharper tips.
Chudoba et al., page 15 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
The reason for the higher indentation hardness is the cushion shape of the indents. Other than
glassy carbon, fused silica shows the largest inwards bowing of the indentation edges after
unloading. This is also the reason for the large radial displacement correction for fused silica
(see paragraph 2.4). The indentation area for the calculation of Vickers hardness is
overestimated, since it is calculated from the diagonal lengths only. Indentation hardness is
therefore a better measure for the plastic properties of fused silica than Vickers hardness and
has a closer connection to the real contact pressure.
The hardness results for sapphire agree well with the reference hardness. The large hardness
Fo
increase at small loads for the blunt Berkovich tip is remarkable. The measurements are fully
elastic up to 10 mN load (fourth data point from left). The hardness value in this load range
rP
represents only the contact pressure and not the plastic properties of the material. Sapphire
reaches here an apparent hardness of 40 GPa, i.e. the theoretical strength of about 1/10th of the
modulus. A reason for this high strength is the high surface quality and low roughness and the
ee
small dimension of the contact area of only 0.23 µm2. The probability for the presence of
flaws around such a small area is low.
rR
W shows a strong hardness increase towards the surface, which is similar for all three tips.
The increase for the rounder Berkovich is smaller. Such a hardness increase towards the
ev
surface can also be observed for the other investigated samples Au-Cu alloy and Ni. This
effect is often described in the literature as indentation size effect (ISE). In fact the hardness
iew
increase cannot be attributed to only one effect. An important feature of the investigated
metals is the modification of the grain structure and the work hardening during the cutting and
polishing procedure. It is known that mechanical preparation modifies the surface to a depth
of some micrometers (see e.g. [25]). Another reason is the required agglomeration of
On
dislocations at the very beginning of the plastic deformation [26]. Gao and Nix have shown
that hardness should increase proportional to 1/√h in crystalline materials (for sharp conical
indenters). Later Qu, Nix et al. [27] have shown that this relation strongly depends on the
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 16 of 34
indenter tip radius. Of course a hardness decrease can be expected in the transition range from
elastoplastic to wholly elastic deformation. The measured hardness gradient in the low depth
range is therefore inevitably a combination of several effects. According to the model of Qu,
the lower hardness increase towards the surface in W for the UMIS data can be attributed to
the larger tip radius. Interestingly the difference was not seen in the nickel.
The hardness increase for the Au-Cu alloy can only be seen for the rounder Berkovich tip
where data are available from smaller depths. At higher depths there is no difference between
the tips and the hardness results agree well with the reference value.
Chudoba et al., page 16 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 17 of 34
The hardness of Si3N4 does not depend on the tip rounding of the Berkovich tip and agrees
with the reference value. However, the hardness, measured with the cube corner indenter, is
smaller and decreases towards the surface. Analogous to the modulus results, the sharper
indenter seems to be more sensitive to the grain size and brittleness of the ceramics. The
opposite effect can be observed for cemented carbide. While the hardness results for both
Berkovich tips agree very well with each other and with the reference value, a hardness
increase is seen for the cube corner indenter. The analysis of this effect is outside the scope of
this investigation.
Fo
The hardness results for glassy carbon show differences between all three indenters,
especially at small loads. The apparent hardness is higher the sharper the tip is. This effect
rP
was not investigated in detail. Possibly it has to do with crack formation in the material as can
be seen in Fig. 3h. More investigations are needed for an understanding of this effect.
The hardness results for Ni do not show significant differences between the indenters. There
ee
is a strong hardness increase towards the surface which reaches nearly two times the bulk
value. The reference value is only reached for the highest loads at a depth of more than 3 µm.
rR
A pile-up effect increases the contact area and the calculated area value from the loaddisplacement curve is too small and leads to a higher apparent hardness.
ev
Summarizing the results of all materials it can be stated, that there is no general hardness
difference seen for the indenters with different tip angle. Hardness results were found to be
iew
similar for metals, glass and sapphire. Significant differences occur for the ceramic materials
and the highly elastic glassy carbon. Some of the differences at small loads can be attributed
to the different tip rounding and to the tip angle.
Hardness is contact pressure under load for a fully developed plastic zone. If such a zone can
On
not develop due to severe cracks or very high elasticity of the material, hardness may no
longer be a reasonable measure for the plastic properties of the material.
Although the maximum strain below the indenter is larger for a smaller face angle, the contact
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
pressure represents an average of the whole strain field. As mentioned in the introduction,
hardness is related to the yield stress Y in simple compression over the constraint factor C.
Tabor [13] has shown that the results for a perfectly plastic solid may be applied for a strainhardening solid if the flow stress for a representative strain εR is used with
ε R ≈ 0.2 ⋅ tan β
(6)
The representative strain is 8% for Vickers and Berkovich indenter and 18% for a cube corner
indenter. This representative strain should decrease if the material is already work hardened
e.g. mechanically polished metal surfaces to a depth of a few micrometers. The constraint
Chudoba et al., page 17 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
factor for metals is then close to three, independent of the indenter angle. Johnson [11,
pp.163] has shown that the dependency of the contact pressure (hardness) on the indenter
angle is further reduced with increasing friction coefficient. Metals have a higher friction
coefficient against diamond than the other investigated materials (typically 0.15 – 0.25). This
may explain why a significant hardness difference between the different indenters was not
observed for the metal samples.
Fo
4. Conclusions
rP
Indentation measurements with Berkovich and cube corner indenters were carried out on eight
different materials with two different commercial nanoindentation instruments over a load
range between 0.3 mN and 500 mN. The hardness and modulus results were compared with
ee
reference values obtained by different methods.
It has been shown that the instrument stiffness can be determined by a comparison of two area
rR
functions obtained using fused silica and sapphire as reference materials. The accuracy of the
area function calculation could be improved by considering additionally the elastic radial
ev
displacement within the contact area. This is especially important if fused silica is used as
reference material because it requires the largest correction.
The modulus results for fused silica, sapphire and glassy carbon show that the unloading fit
iew
range has a much bigger influence on the results for the cube corner indenter than for the
Berkovich tip. Fused silica and glassy carbon have a characteristic kink in the unloading
curve independent of the maximum load. Therefore the unloading curves of a cube corner
On
indenter should be carefully inspected to ensure smooth curvature and the unloading fit range
has to be chosen warily to enable a reliable comparison of the results. A fit range of less than
50% is recommended.
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 18 of 34
A comparison of the modulus results has shown that there is no significant difference between
the results from different instruments or between cube corner and Berkovich indenters. This
confirms that the same calibration procedures can be applied, independent of the indenter
shape. Further, the indentation modulus results agree well with the reference values obtained
by other methods. A good agreement with the Young’s modulus reference value of the very
stiff cemented carbide confirms the correctness of the calculated instrument compliance, since
an error in the compliance correction would have the highest influence on results from this
Chudoba et al., page 18 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 19 of 34
material. Cracks in sapphire and fused silica, which evolve mainly in the upper load range,
had no noticeable influence on the modulus results.
The hardness comparison has shown that for the majority of the investigated materials, there
is no fundamental hardness difference for indenters with different tip angles. This means that
the contact pressure below the indenter is approximately equal after development of the full
plastic zone. The difference between Berkovich and cube corner indenter increases with
decreasing contact depth. Some of the differences at small loads can be attributed to the
different tip rounding. Materials like silicon nitride, cemented carbide or glassy carbon,
Fo
however, show a clear deviation from the hardness reference value for the cube corner
indenter although the modulus difference is not so much pronounced. Deviations are equally
rP
possible above as below the reference value. The reasons for the deviations were not studied
in detail. In the case of silicon nitride the brittleness of the material seems to result in a microcrack enhancement for the sharper tip and therefore in a hardness decrease. For cemented
ee
carbide the hard WC particles in the softer cobalt matrix seem to determine the total response
of the compound more for a sharper tip. The slip-line theory (see [9]) of metal forming cannot
be applied for these materials.
rR
Generally the scatter of the results is larger for the sharper cube corner tip. Therefore
ev
Berkovich indenters should be preferred for comparable hardness and modulus results. Cube
corner indenters require a careful calibration but can then be used, for instance as an
iew
additional tip for the evaluation of the stress-strain relation of plastic materials.
Acknowledgements
On
The authors thank M. Griepentrog and A. Dück from the Federal Institute of Materials
Research and Testing (BAM), Berlin, Germany, for the hardness and modulus reference
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
measurements. G. Bürki and S. Stauss from Empa in Thun are acknowledged for their help
with the in-situ SEM indentations.
This research activity was financially supported by the Swiss Office of Education and Science
(OFES): European project: ROBOSEM – Growth: G1RD-CT-2002-00675.
Chudoba et al., page 19 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
Table 1:
Hardness and modulus results obtained with different tips and instruments as an average of all
measurements with different maximum load. ∆H is the standard deviation of 10 measurements
at 500mN for the reference measurements. For the indentation measurements ∆H and ∆E are
the standard deviations of the results of all loads for one sample without including the scatter
of the ten measurements averaged at each load. Hardness results below 40 nm indentation
depth are not included in the average.
rP
Fo
Reference
7.5
23.6
5.39
2.53
19.5
30.1
3.9
1.63
E
H
8.8
22.5
5.9
2.65
19.0
29.1
3.8
2.7
∆H
0.3
2.2
1.0
0.22
1.0
1.8
0.24
0.4
∆E
1.9
10.2
25.5
4.2
6.3
24.5
0.2
19.3
H
XP Cube corner
∆H
0.6
1.3
1.6
0.05
0.6
1.4
0.3
0.7
H
8.1
21.6
6.5
2.32
17.7
28.4
4.6
2.3
8.4
22.2
6.4
2.24
16.1
36.4
4.9
2.1
∆H
0.7
1.2
1.7
0.15
2.3
5.4
0.7
0.4
E
71.5
415.9
414.6
90.5
293.4
621.4
27.2
226.6
∆E
0.8
7.3
33.8
1.1
3.7
16.4
1.1
9.2
E
72.2
401.5
394.4
100.6
260.2
614.6
25.5
241.0
∆E
0.8
19.0
36.2
4.6
22.2
38.1
1.2
5.5
ly
On
E
71.5
430.8
413.8
98.9
299.9
642.4
25.9
227.0
XP Berkovich
iew
72
420
410
93
304
654
27
200
∆H
0.1
1.1
0.17
0.05
2.3
1.5
0.4
0.04
ev
E-Modulus in
Gpa
Fused silica
Sapphire
W
Au-Cu alloy
Si3N4
Cem. Carbide
Glassy carbon
Ni
H
UMIS Berkovich
rR
Hardness in
GPa
Fused silica
Sapphire
W
Au-Cu alloy
Si3N4
Cem. Carbide
Glassy carbon
Ni
ee
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 20 of 34
Chudoba et al., page 20 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 21 of 34
List of Figure Captions
Figure 1:
Indentation modulus of (a) fused silica 500mN maximal load; (b) sapphire 100mN maximal
load, and, (c) glassy carbon 500 mN maximal load, determined as a function of the unloading
fit range for a Berkovich (squares) and a cube corner (triangles) tip. The measurements have
been done with a MTS Nanoindenter XP. The horizontal solid line indicates the reference
Fo
modulus value.
Figure 2:
rP
Contact depth dependent modulus results for the investigated materials measured with UMIS
Berkovich (circles), XP Berkovich (squares), and XP cube corner (triangles). The horizontal
ee
solid lines indicate the reference modulus values.
Figure 3:
rR
In-situ SEM indentations in sapphire (a, b) at 981 mN, Ni (c, d) at 343 mN, W (e, f) at
ev
981 mN and glassy carbon (g, h) at 343 mN using a cube corner with the SEM-Microindenter.
The left column shows the indented material close to or at maximum load and the right
column the residual indent. Note the pronounced pile-up around the indent for Ni and W.
Figure 4:
iew
Contact depth dependent indentation hardness for all investigated materials measured with
On
UMIS Berkovich (circles), XP Berkovich (squares), and XP cube corner (triangles). The
horizontal solid lines indicate the reference hardness values.
ly
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
Chudoba et al., page 21 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
a)
Indentation modulus (GPa)
80
Cube corner
Berkovich
Reference
75
70
Fo
10
30
425
ev
375
20
30
40
50
60
70
80
Fit range from F max(%)
100
ly
Cube corner
Berkovich
Reference
90
On
25
iew
Indentation modulus (GPa)
100
Cube corner
Berkovich
Reference
30
c)
90
rR
400
10
40
50
60
70
80
Fit range from F max (%)
ee
b)
Indentation modulus (GPa)
450
20
rP
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 22 of 34
20
10
20
30
40
50
60
70
80
Fit range from F max (%)
90
100
Fig. 1
Chudoba et al., page 22 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 23 of 34
iew
ev
rR
ee
rP
Fo
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
ly
On
Fig. 2
Chudoba et al., page 23 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
b)
a)
2 µm
c)
d)
rP
10 µm
10 µm
f)
ev
rR
ee
e)
2 µm
Fo
5 µm
10 µm
g)
h)
5 µm
5 µm
iew
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 24 of 34
Fig. 3
Chudoba et al., page 24 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 25 of 34
iew
ev
rR
ee
rP
Fo
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
ly
On
Fig. 4
Chudoba et al., page 25 of 27
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
References
[1]
[2]
[3]
[4]
[17]
[18]
[19]
[20]
ly
[16]
On
[15]
iew
[13]
[14]
ev
[8]
[9]
[10]
[11]
[12]
rR
[7]
ee
[6]
rP
[5]
C.B. Ponton, R.D. Rawlings, Vickers indentation fracture toughness test, Part 1 and 2,
Mat. Sci. and Technol. 5, 865-976 (1989).
D.S. Harding, W.C. Oliver, G. M. Pharr, Cracking during nanoindentation and its use in
the measurement of fracture toughness, Mat. Res. Soc. Symp. Proc. 356, 663-668
(1995).
S. Stauss , P. Schwaller , J.-L. Bucaille , R. Rabe , L. Rohr , J. Michler, E. Blank,
Determining the stress–strain behaviour of small devices by nanoindentation in
combination with inverse methods, Microelectronic Engineering 67–68, 818–825
(2003).
J.L. Bucaille, S. Stauss, E. Felder, J. Michler, Determination of plastic properties of
metals by instrumented indentation using different sharp indenters, Acta Materialia 51,
1663 (2003).
J.-L. Bucaille, S. Stauss , P. Schwaller , J. Michler, A new technique to determine the
elastoplastic properties of thin metallic films using sharp indenters, Thin Solid Films
447-448, 239-245 (2004).
N. Chollacoop, M. Dao, S. Suresh, Depth-sensing instrumented indentation with dual
sharp indenters, Acta Materialia 51, 3713–3729 (2003).
J. S. Field, M. V. Swain, R. D. Dukino, Determination of fracture toughness from the
extra penetration produced by indentation-induced pop-in, J. Mater. Res. 18, 1412-1419
(2003).
W.C. Oliver, G.M. Pharr, J. Mater. Res. 7, 1564-1583 (1992).
R. Hill, The Mathematical Theory of Plasticity, (Clarendon Press, Oxford, 1950).
D.M. Marsh, Plastic flow in glass, Proc. R. Soc. London, A279, 420-435 (1964).
K.L. Johnson, Contact Mechanics, (Cambridge University Press, Cambridge, 1985).
J. Zhang, M. Sakai, Geometrical effect of pyramidal indenters on the elastoplastic
contact behaviors of ceramics and metals, Materials Science and Engineering A 381, 6270 (2004)
D. Tabor, Hardness of Metals, Oxford, University Press (1951)
ISO 14577, Metallic materials – Instrumented indentation test for hardness and
materials parameters, 2002, ISO Central Secretariat, 1211 Geneva 20, Switzerland.
R. Rabe, J.-M. Breguet, P. Schwaller, S. Stauss, F.-J. Haug, J. Patscheider and J.
Michler, Observation of fracture and plastic deformation during indentation and
scratching inside the scanning electron microscope, Thin Solid Films, 469-470, 206-213
(2004)
European project, EC Contract G6RD-CT2000-00418 "Certified Reference Materials
for Depth Sensing Indentation Instruments 'DESIRED'
European project: Determination of Hardness and Modulus of Thin Films and Coatings
by Nanoindentation (INDICOAT), contract no. SMT4-CT98-2249, NPL Report
MATC(A) 24, May 2001.
Tables of Physical and Chemical Constants, Ed. W. G.C. Kaye, T.H. Laby, (Longman,
London, 1995).
T. Chudoba, M. Griepentrog, A. Dück, D. Schneider, F. Richter, Young’s modulus
measurements on ultra-thin coatings, J. Mater. Res. 19, 301-314 (2004).
W.C. Oliver, G.M. Pharr, Measurement of hardness and elastic modulus by
instrumented indentation: Advances in understanding and refinements to methodology,
J. Mater. Res. 19, 3-20 (2004).
Fo
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 26 of 34
Chudoba et al., page 26 of 27
http://mc.manuscriptcentral.com/pm-pml
Page 27 of 34
[21] R.B. King, Elastic analysis of some punch problems for a layered medium, Int. J. Solids
Structures 23 no. 12, 1657-1664 (1987).
[22] M.T. Hendrix, The use of shape correction factors for elastic indentation measurements,
J. Mater. Res. 10, 255-258 (1995).
[23] J. B. Wachtman Jr., W.E. Tefft, D.G. Lam Jr., R.P. Stinchfield, Elastic constants of
synthetic single crystal corundum at room temperature. J. Res. Nat. Bur. Stand. A.
Physics and Chemistry 64A, 213–228 (1960).
[24] J. J. Vlassak, M. Ciavarella, J. R. Barber, X. Wang, The indentation modulus of
elastically anisotropic materials for indenters of arbitrary shape, Journal of the
Mechanics and Physics of Solids, 51, 1701 – 1721 (2003).
[25] Y. Liu, A.H.W. Ngan, Depth Dependence of Hardness in Copper Single Crystals
Measured by Nanoindentation, Scripta Mater 44, 237-241 (2001).
[26] W.D. Nix, H. Gao, Indentation size effects in crystalline materials: a law for strain
gradient plasticity, J. Mech. Phys. Solids 46, 411-425 (1998).
[27] S. Qu, Y, Huang, W.D. Nix, H. Jiang, F. Zhang, K.C. Hwang, Indenter tip radius effect
an the Nix-Gao relation in micro- und nanoindentation hardness experiments, J. Mater
Res. 19, 3423-3434 (2004)
iew
ev
rR
ee
rP
Fo
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
Chudoba et al., page 27 of 27
http://mc.manuscriptcentral.com/pm-pml
rP
Philosophical Magazine & Philosophical MagazinePage
Letters
28 of 34
80
Indentation modulus (GPa)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
ee
Cube corner
Berkovich
Reference
rR
75
ev
iew
70
10
20
30
40
50
60
70
80
Fit range from Fmax (%)
http://mc.manuscriptcentral.com/pm-pml
On
90
100
ly
Page 29 of 34
ev
rR
ee
rP
Fo
Figure 1(a)
300x232mm (150 x 150 DPI)
w
ie
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
http://mc.manuscriptcentral.com/pm-pml
or
Pe
Philosophical Magazine & Philosophical Magazine Page
Letters
30 of 34
450
Indentation modulus (GPa)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
425
er
vie
400
w
375
10
Re
20
30
Cube corner
Berkovich
Reference
40
50
60
70
80
Fit range from Fmax(%)
http://mc.manuscriptcentral.com/pm-pml
On
90
100
ly
rP
Page Philosophical
31 of 34
Magazine & Philosophical Magazine Letters
30
Indentation modulus (GPa)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
ee
rR
25
ev
iew
20
10
Cube corner
Berkovich
Reference
20
30
40
50
60
70
80
Fit range from Fmax (%)
http://mc.manuscriptcentral.com/pm-pml
On
90
100
ly
Philosophical Magazine & Philosophical Magazine Letters
w
ie
ev
rR
ee
rP
Fo
Figure 2
170x241mm (150 x 150 DPI)
http://mc.manuscriptcentral.com/pm-pml
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 32 of 34
Page 33 of 34
w
ie
ev
rR
ee
rP
Fo
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philosophical Magazine & Philosophical Magazine Letters
Figure 3
131x190mm (150 x 150 DPI)
http://mc.manuscriptcentral.com/pm-pml
Philosophical Magazine & Philosophical Magazine Letters
w
ie
ev
rR
ee
rP
Fo
ly
On
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4
163x236mm (150 x 150 DPI)
http://mc.manuscriptcentral.com/pm-pml
Page 34 of 34